See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/248597446

Reliability and Validity of the Chinese Behavioural Inattention Test–Hong

Kong Version (CBIT-HK) for Patients with Stroke and Unilateral Neglect

Article in Hong Kong Journal of Occupational Therapy · January 2007

DOI: 10.1016/S1569-1861(07)70004-9

CITATIONS READS

7 494

7 authors, including:

Kenneth N K Fong Marko Chan

The Hong Kong Polytechnic University 12 PUBLICATIONS 216 CITATIONS
147 PUBLICATIONS 1,504 CITATIONS
SEE PROFILE
SEE PROFILE

Bill Chan
Hong Kong Hospital Authority
1 PUBLICATION 7 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Learning environment and approaches to studying among occupational therapy students View project

All content following this page was uploaded by Kenneth N K Fong on 20 May 2015.

The user has requested enhancement of the downloaded file.

 HKJOT 2007;17(1):23–33

RELIABILITY AND VALIDITY OF THE CHINESE

BEHAVIOURAL INATTENTION TEST–HONG KONG
VERSION (CBIT-HK) FOR PATIENTS WITH
STROKE AND UNILATERAL NEGLECT
Kenneth N.K. Fong1, Marko K.L. Chan2, Bill Y.B. Chan2, Peggie P.K. Ng2,
Mei Ling Fung2, May H.M. Tsang2 and Kathy K.Y. Chow2

Objective: One of the explanations for the wide range in the reported incidence of unilateral neglect is
probably the lack of agreement regarding evaluation methods. The aim of this study was to investigate the
reliability and validity of the newly developed Chinese Behavioural Inattention Test–Hong Kong version
(CBIT-HK) for research purposes, and for patients with stroke and unilateral neglect in Hong Kong.
Methods: Fourteen consecutive subacute patients with stroke were selected to determine the external
reliabilities and new cutoff scores, and 54 subacute right hemisphere patients were given evaluations
including the CBIT-HK, a clock drawing test, and a functional assessment battery to find out the inter-
nal reliability and concurrent validity of the test battery.
Results: The results indicated that the interrater, intrarater, and parallel-form reliability of the test were
high [ICC (2,k) = 0.75 to 0.99]. The findings of low cutoffs and sensitivities for two language-related
items in the Behavioural Subtest: “Article Reading” and “Address and Sentence Copying”, suggested
that these items should be used with caution. The significant relationship between the Conventional
Subtest with the clock drawing test (r = 0.34) and the functional assessment battery (r = 0.47) suggested
that the Conventional Subtest, especially the “Star Cancellation” item, rather than the Behavioural
Subtest, could be used as a single screening test or item for neglect in the initial assessment of patients
with stroke by occupational therapists in daily practice.
Conclusion: The reliability and validity of the CBIT-HK were established. The results of this study
should aid in making the CBIT-HK, designed to assess the unilateral neglect or visual inattention of
patients with stroke and unilateral neglect, a tool for clinical application in occupational therapy as well
as for future study in Hong Kong.

KEY WORDS: Chinese Behavioural Inattention Test • Stroke • Unilateral neglect

Introduction hemisphere of the brain, but whose deficit cannot be attrib-

uted to sensory or motor impairment (Ladavas, Menghini &
Unilateral neglect (sometimes known as hemispatial agnosia) Umilta, 1994). Neglect is commonly accepted as a behavioural
is a common phenomenon in patients with hemiplegia that attentional deficit. However, different components of attention
occurs in those who do not respond to, identify or orient to- being affected are emphasized by theorists, such as orienting
wards meaningful stimuli which are contralateral to the lesion (Heilman et al., 1985; Kinsbourne, 1987), selective attention

1
Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, and 2Occupational Therapy Department, Kowloon Hospital,
Hospital Authority, Hong Kong SAR, China.
Reprint requests and correspondence to: Dr. Kenneth N.K. Fong, Department of Rehabilitation Sciences, The Hong Kong Polytechnic
University, Hung Hom, Hong Kong SAR, China.
E-mail: [email protected]

Hong Kong Journal of Occupational Therapy 23

©2007 Elsevier. All rights reserved.
Kenneth N.K. Fong, et al HKJOT 2007;17(1)

(Kaplan et al., 1991; Rapcsak et al., 1989), shifting attention Subtest batteries or their items can be used individually as
(Posner et al., 1987); and automatic attention (Riddoch & each contains an individual cutoff score for the norm. However,
Humphreys, 1983). Halligan & Marshall (1992) suggest that most of the items that are found in the Behavioural Subtest
there is no unified theoretical framework that explains under- include both linguistic and cultural relevant items that have to
lying mechanism of neglect. be changed accordingly when it is to be used in the Chinese
Unilateral neglect is most typically seen to be associated with population in Hong Kong. The first prototyped Chinese version
stroke following a lesion of the right brain hemisphere, although of the BIT was developed by a group of physicians in Tongji
some degree of visual spatial malfunction may also exist with Hospital in Wuhan, China, in 2001. However, that version can-
left hemisphere lesions (Wilson, Cockburn & Halligan, 1987a). not be used in Hong Kong owing to the contextual differences
The reported incidence of unilateral neglect in patients with between the two places such as food or money (coins) used in
right hemisphere lesion varies widely from 12% to 90%, in part the instrument. Therefore, the purpose of this study was to
because of the cross-study differences (Massironi et al., 1988; investigate the reliability and validity of the newly developed
Sunderland, Derick & Hewer, 1987). One of the explanations Chinese Behavioural Inattention Test–Hong Kong version
is the lack of objective assessment of unilateral neglect. There (CBIT-HK), for research purposes, with the linguistic part
was no consensus on evaluating unilateral neglect and it might built up on top of the prototyped Chinese version of the BIT in
be consistent among some patients despite lack of objective Wuhan, and to determine new cutoff scores for the Behavioural
behavioural observations (Wilson, Cockburn & Halligan, 1987). Subtest of the CBIT-HK, if any, for normality and for patients
Severe unilateral neglect is considered to have serious effects on with stroke and unilateral neglect in Hong Kong.
rehabilitation and functional recovery (Feigenson, McCarthy &
Greenberg, 1977; Kinsella & Ford, 1980; Denes, Semenza & Methods
Stoppa, 1982; Kotila, Niemi & Laaksonen, 1986; Sunderland
et al., 1987; Calvanio, Levine & Petrone, 1993; Katz et al., Subjects
1999). It is therefore important to develop standardized and This study followed the method of selection and sample size
objective methods of evaluating or detecting unilateral neglect in the manual of the original standardized version of the test
so that appropriate intervention can be provided to remediate battery (Wilson, Cockburn & Halligan, 1987b). The study was
the impairment in daily clinical practice. divided into two parts. Fourteen patients with subacute stroke,
To assess unilateral neglect, traditional evaluation using who received occupational therapy in a rehabilitation hospital
letter cancellation, clock face drawing, and figure copying are in September to November 2002, was selected by convenient
either fragmented or cannot quantify the degrees of impairment sampling to join part I of the study. The inclusion criteria for
in relation to severity of disability in performing common daily the subjects were: (1) > 18 years old; (2) suffering from stroke
activities. The Behavioural Inattention Test (BIT) was developed (haemorrhagic or ischaemic) with either hemiplegic side; (3) in-
as a criterion-referenced test for unilateral neglect or visual or outpatient in the occupational therapy department; (4) with or
inattention of patients suffering from stroke or brain injuries without unilateral neglect/visual inattention; (5) able to under-
(Wilson, Cockburn & Halligan, 1987). It has been used exten- stand instructions of the test battery; and (6) able to read sim-
sively in past research studies on unilateral neglect and some ple sentences and draw simple pictures with either hand. The
also used it for detecting visual inattention (Wilson, Cockburn & mean standard deviation (SD) age of the subjects was 63.9
Halligan, 1987a; Robertson et al., 1990; Katz et al., 1999). The (13.5). There were 11 (79%) inpatients and three (21%) out-
BIT is divided into two parts: the Conventional Subtest and patients; three (21%) were right hemiplegic and 11 (79%) had
the Behavioural Subtest. The Conventional Subtest includes six left hemiplegia. Five (36%) were detected to have unilateral
items: Line Crossing, Letter Cancellation, Star Cancellation, neglect by the case therapist in daily behavioural observations
Figure and Shape Copying, Line Bisection, and Representative and nine (64%) did not. The average (SD) time between stroke
Drawing. The Behavioural Subtest includes: Picture Scanning, onset and admission to the programme was 129 (185.3) days.
Telephone Dialling, Clock Telling, Menu Reading, Article The mean (SD) of formal education was 3.2 (2.3) years. They
Reading, Map Reading, Coin Sorting, Address and Sentence were used to establish the interrater, intrarater, and parallel-
Copying, and Card Sorting. The total score for the Conven- form reliability as well as the new cutoff scores for individual
tional Subtest is 146, and 81 for the Behavioural Subtest, mak- items of the Behavioural Subtest of the CBIT (Table 1).
ing a total score of 227. Lower scores indicate more neglect or Fifty-four subjects, who were subacute inpatients with stroke
inattention. There are two parallel versions for the Behavioural and unilateral neglect, were then recruited consecutively in the
Subtest: versions A and B. Both the Conventional or Behavioural same hospital to participate in the study between November

24 Hong Kong Journal of Occupational Therapy

 RELIABILITY AND VALIDITY OF THE CBIT-HK

2003 and July 2005 in a rehabilitation hospital in Hong Kong of each character as well as the length of each sentence in
to determine internal reliability and concurrent validity (Part Chinese to make sure that the format of the presentation matched
II). Eligibility criteria included: (1) stroke (haemorrhagic or the original English sentences. For the other items, the investi-
ischaemic) confirmed by medical record; (2) neurological rep- gators changed the contents of the items “Picture Scanning”
resentation compatible with unilateral right lesion involve- (Figure 3), “Telephone Dialling”, “Telling and Setting the Time”,
ment (i.e. left hemiplegic); (3) right handedness; (4) < 8 weeks “Menu Reading”, “Coin Sorting”, and “Map Navigation”, with
since onset of stroke at study entry; (5) a score of 15 on the reference to the Hong Kong cultural context. They included:
Glasgow Coma Scale during recruitment; and (6) unilateral (1) changing Picture I of “Picture Scanning” to culturally rel-
neglect/visual inattention as determined by case therapists in evant food in Hong Kong; (2) increasing the dial number by
daily observations. Subjects were excluded if they had severe one more digit for “Telephone Dialling” to make eight digits
dysphasia (either expressive or comprehensive) that restricted as commonly used in Hong Kong; (3) adding numbers to the
communication, or significant impairment in visual acuity
caused by cataracts, diabetic retinopathy, glaucoma, or hemi-
Table 1. Summary of sample characteristics in Parts I and II
anopia. The mean (SD) age of the subjects was 69.7 (10.5)
of the study
years. Thirty-four (63%) subjects were men, and 20 (27%)
Part I Part II
were women. The average (SD) time between stroke onset and
(n = 14) (n = 54)
admission to the programme was 11.9 (7.3) days. The average
length of stay was 44 (18.1) days. The subjects had 3.2 (2.5) Gender, n (%)
Male 11 (79) 34 (63)
years of formal schooling (Table 1).
Female 3 (21) 20 (27)
The prototyped CBIT-HK was developed and first exam-
Medical history
ined by an expert panel composed of occupational therapists
Period from onset*, mean (SD) 129 (185.3) 11.9 (7.3)
with extensive experience working with patients with stroke. Right hemiplegic, n (%) 3 (21) 0
The Chinese translations of the items “Article Reading” Left hemiplegic, n (%) 11 (79) 54 (100)
(Figure 1) and “Address and Sentence Copying” (Figure 2) were With neglect, n (%) 5 (36) 54 (100)
Without neglect, n (%) 9 (64) 0
based on the version developed by the Tongji Hospital (2001) in
Inpatient, n (%) 11 (79) 54 (100)
Wuhan. The investigators changed the simplified Chinese to tra- Outpatient, n (%) 3 (21) 0
ditional Chinese characters, and the panel identified the Chinese
Social history
words or phrases that did not reflect the semantic meaning Age, yr, mean (SD) 63.9 (13.5) 69.7 (10.5)
with reference to the Hong Kong cultural context. The results Yrs of education, mean (SD) 3.2 (2.3) 3.2 (2.5)
revealed that there was no need to change any of the linguistic *Period from onset = the average time between stroke onset and
content. The investigators then measured accurately the height assessment in the study (days). SD = standard deviation.

Figure 1. Article reading of the Behavioural Subtest of the CBIT-HK.

Hong Kong Journal of Occupational Therapy 25

Kenneth N.K. Fong, et al HKJOT 2007;17(1)

Figure 2. Sentence copying of the Behavioural Subtest of the CBIT-HK.

parallel-form reliability, and intrarater reliability, and establish-

ment of new cutoffs for individual subtests of the Behavioural
Subtest of the CBIT-HK on a single time occasion. For inter-
rater reliability, two examiners assessed the subject using the
Conventional Subtests and Version A of the Behavioural
Subtest of the CBIT-HK simultaneously during the initial
assessment and rated separately. The second examiner read the
instructions to each subject. The first examiner was blinded
for the subject recruitment. For parallel-form reliability, since
there were two parallel forms for Version B but a single form
only for Version A, the first examiner assessed the same subject
again, with version B of the CBIT-HK, after 3 consecutive
days. For intrarater (test–retest) reliability, the first examiner
assessed the same subject again with the Conventional
Subtest and Version A of the Behavioural Subtest after a time
Figure 3. Picture scanning of the Behavioural Subtest of the
CBIT-HK. interval of 1 week. Part II of the study involved item analysis of
the CBIT-HK and investigation of its concurrent validity with
other functional and cognitive measures as applied to patients
clock face in “Telling and Setting the Time”; (4) adapting “Menu with stroke and unilateral neglect on a single time occasion.
Reading” with Chinese characters for common foods found in For internal consistency, item analysis was investigated to deter-
restaurants in Hong Kong; (5) changing “Coin Sorting” with mine internal reliability. Concurrent validity was established
the values of common Hong Kong coins. For the Conventional by correlating the CBIT-HK with: (1) a criterion functional
Subtest of the CBIT-HK, no local adaptation was determined to parameter—the Functional Independence Measure (FIM) (UDS,
be necessary for application in the local context as the test items 1994); and (2) cognitive assessments including the Mini-
were culture-free. The Chinese test manual and the recording Mental State Examination (MMSE) (Chiu et al., 1994), and the
form were amended from the Wuhan version, which was a direct Clock Drawing Test (CDT) (Watson, Arfken & Birge, 1993).
English-to-Chinese translation of the standardized instruc- For both Parts I and II of the study, examiners attended a
tions for test administration of the test battery (Appendix). 2-hour training session on the administration of the test
beforehand. The session included practical training on the use
Research Design of the Chinese prototype of the BIT and its scoring method.
The study for reliability and concurrent validity consisted of The measures were administered in a quiet room in the occu-
two parts. Part I was the investigation of interrater reliability, pational therapy department during the therapy session. Written

26 Hong Kong Journal of Occupational Therapy

 RELIABILITY AND VALIDITY OF THE CBIT-HK

and informed consent was sought from all participants prior version 12.0 (SPSS Inc., Chicago, IL, USA). Intraclass corre-
to data collection. If the subjects were not able to give their lation [ICC (2,k)] two-way absolute agreement was used to
own consent mentally, health care staff or investigators represent the interrater, parallel-form, and test–retest reliabili-
explained the study procedure to obtain the support and ties. Item analysis was conducted to investigate Cronbach’s α
understanding from their next of kin or relatives. All subjects and item–total correlation. Sensitivity and specificity using the
were told that they had the right to withdraw from the study RO curve was used to determine the new cutoff scores for the
at any time. All data were kept confidential throughout the Behavioural Subtest of the CBIT-HK with comparison with
study period. the original scores.

Instruments Results
Four instruments were used for the assessment of subjects in
part II of the study. These were the CBIT-HK developed in this For part I of the study, ICC (2,k) for interrater reliability was
study, the CDT (Watson, Arfken & Birge, 1993), the MMSE 0.99 for the Conventional Subtest, 0.99 for the Behavioural
(Chiu et al., 1994), and the FIM (UDS, 1994). The CDT was Subtest, and 0.99 for the CBIT-HK total (n = 14) (Table 2). The
used to quantify the extent of neglect in clock drawing rather results of the parallel-form reliability of Versions A and B of
than scoring 1 or 0 in the CBIT-HK. It has been used for neu- the Behavioural Subtest indicated that the reliability was high
ropsychological screening of visuo-constructional abilities [ICC (2,k) = 0.95] (n = 10) (Table 2). For intrarater reliability,
and the Watson system is most likely to be the best predictor the results were found to be satisfactory, with ICC (2,k) rang-
of right hemispheric strokes among several clock drawing sys- ing from 0.75 to 0.84 (n = 7) (Table 2).
tems due to its focus on spatial placement of numbers in the Because there was no change in the content of the Conven-
scoring criteria. The total score ranges from 1 to 7, with lower tional Subtest, we used the cutoff scores obtained from the
scores indicating more neglect and inattention. The standard- original British sample (Wilson, Cockburn & Halligan, 1987b).
ized Chinese version of the MMSE is commonly used by However, almost all the content in the Behavioural Subtest
occupational therapists as a screening test for cognitive per- had been revised during the validation process, so the cutoff
formance in Hong Kong. It takes about 15 minutes to adminis- scores for all the items in this subtest were revised. The optimal
ter and ends with a total score of 30 for optimal performance. cutoff scores were determined by finding which item had
The cutoff point of 19 can discriminate between normal adults a maximum sensitivity and specificity using the RO curves
and subjects with cognitive impairment. The FIM is used to (Table 3). Results indicated that the optimal cutoff scores for
measure the degree of independence and assistance needed most of the items, except “Article Reading”, “Address and
for activities of daily living. The FIM consists of 18 items that Sentence Copying”, and “Card Sorting”, were most sensitive to
assess a broad range of activities of daily living, including two correctly classify 80% of having neglect in subjects with uni-
unidimensional subscales comprising 13 items assessing lateral neglect, and 56–89% of the subjects without neglect were
motor function, i.e. the motor measure (FIM-MM), and five classified in the sample as indicated in the relatively moderate
items assessing cognitive function, i.e. the cognitive measure to high specificity. Low cutoff scores were suggested for two
(FIM-CM). The items are therapist-rated based on the patient’s language-related items, i.e. “Article Reading” and “Address and
performance according to a 7-point Likert scale. A score of 1 Sentence Copying”, with a sensitivity and specificity of 0.60
means total assistance is needed while a score of 7 means the and 0.78, and 0.40 and 0.78, respectively (Table 3).
patient is completely independent. The four subtotal scores of Results of internal consistency indicated that both the
the motor measure according to the functional activities are: Conventional Subtest total score and the Behavioural Subtest
total subscores of self-care, sphincter control, transfer, and total score had the highest item–total correlation (0.970 and
locomotion, and the five subtotal scores of the cognitive mea- 0.924 respectively) and the lowest Cronbach’s α values if the
sures are: total subscores of comprehension, expression, social item was deleted from the scale (0.805 and 0.815 respectively)
interaction, problem-solving, and memory. (Table 4), i.e. they had the highest internal reliability with the
whole test battery, with the Conventional Subtest being slightly
Statistical Analysis more reliable than the Behavioural Subtest. Results also indi-
All data were treated as continuous data and analysed using cated that the “Star Cancellation” item was the most reliable
parametric statistical methods. The total scores, the conventional individual item apart from the two subtest scores, 0.882 for
and behavioural subscores of the CBIT-HK, as well as individual item–total correlation and 0.823 for Cronbach’s α if the item
item scores were analysed using the statistical software SPSS was deleted from the whole battery (Table 4).

Hong Kong Journal of Occupational Therapy 27

Kenneth N.K. Fong, et al HKJOT 2007;17(1)

Table 2. Interrater, parallel-form and intrarater reliabilities of CBIT-HK

Interrater reliability Parallel-form reliability Intrarater reliability
ICC (2,k)
(n = 14) (n = 10) (n = 7)

Conventional Subtest total 0.99 0.82

Behavioural Subtest total 0.99 0.95 0.75
CBIT total 0.99 0.84

Table 3. Sensitivity, specificity, and the new cutoff score of the Behavioural Subtest of the CBIT-HK (n = 15)
Maximum Original cutoff New cutoff
Items Sensitivity (%) Specificity (%)
score score score

Picture scanning 9 5 5 0.80 0.67

Telephone dialling 9 7 7 0.80 0.89
Menu reading 9 8 8 0.80 0.67
Article reading 9 8 5 0.60 0.78
Telling and setting the time 9 8 8 0.80 0.56
Coin sorting 9 8 8 0.80 0.89
Address and sentence copying 9 7 4 0.40 0.78
Map navigation 9 8 8 0.80 0.89
Card sorting 9 8 8 0.40 0.78
Behavioural Subtest total 81 67 61 0.80 0.78

Table 4. Item statistics of the CBIT-HK (n = 54)

Item–total Squared multiple Cronbach’s α* if
Items Mean SD
correlation correlation item is deleted

Line crossing 24.11 11.77 0.852 0.990 0.830

Letter cancellation 12.22 11.18 0.843 0.988 0.831
Star cancellation 25.02 15.60 0.882 0.993 0.823
Figure copying 0.35 0.68 0.473 0.661 0.852
Line bisection 2.39 2.57 0.307 0.840 0.851
Representative drawing 0.56 0.93 0.380 0.640 0.852
Conventional Subtest total 64.44 36.23 0.970 1.000 0.805
Picture scanning 2.24 2.36 0.691 0.986 0.848
Telephone dialling 3.04 3.42 0.703 0.993 0.847
Menu reading 4.17 3.94 0.750 0.995 0.845
Article reading 1.20 2.84 0.560 0.992 0.849
Telling and setting the time 5.70 3.27 0.716 0.993 0.847
Coin sorting 2.37 2.67 0.713 0.986 0.848
Address and sentence copying 1.30 3.02 0.570 0.991 0.848
Map navigation 5.93 3.45 0.715 0.993 0.846
Card sorting 2.61 3.05 0.658 0.992 0.848
Behavioural Subtest total 28.61 20.60 0.924 1.000 0.815
CBIT total 93.07 54.96 1.000 1.000 0.856
*Cronbach’s α for the CBIT-HK = 0.849. SD = standard deviation.

For part II of the study, results of Spearman’s rho correla- similar and had a mild to moderate significant relationship
tion coefficient indicated that there was a higher significant with the self-care, transfer, and social cognition subscales, the
relationship between the total scores of the CBIT and the total scores of the Motor and Social Cognition subscales, and
Conventional Subtest (r = 0.98), as compared with the the total FIM (r = 0.35 to 0.47), with the coefficients slightly
Behavioural Subtest (r = 0.95). The pattern of both subtests was higher for the Conventional Subtest than for the Behavioural

28 Hong Kong Journal of Occupational Therapy

 RELIABILITY AND VALIDITY OF THE CBIT-HK

Subtest. There was no significant relationship between CBIT-HK deviation (crossover effect) was found in some patients, which
and its subtests with the sphincter, locomotion and communi- made the task not sensitive enough to be the screening test for
cation subscales of the FIM. Particularly noteworthy is the neglect (Azouvi et al., 2003). It is not a surprise that the can-
finding that the Conventional Subtest and total CBIT-HK had cellation tasks are perhaps the most widely used and sensitive
a significant relationship with the CDT (r = 0.34), but not for the tasks for neglect, which can correctly classify about 50% of
Behavioural Subtest (Table 5). In addition, only total CBIT-HK neglect (Azouvi et al., 2003). Line cancellation was easier for
and the Behavioural Subtest related significantly with educa- patients with neglect as the task had only one target item (no
tional level (r = 0.27 and 0.30 respectively). target embedded within an array of many different types of
distractor items); patients with right hemispheric lesion can
Discussion easily start to search from the right of the array (Parton et al.,
2004). Thus, given that Chinese patients had difficulty in doing
The aim of this study was to investigate the reliability and valid- English alphabetical letter cancellation, it is suggested that
ity of the new CBIT-HK for research purposes. The present the “Star Cancellation” task can be used as a single screening
findings are consistent with the results reported in a previous item in the initial assessment of patients with stroke by any
study, i.e. that both the interrater and parallel-form reliabili- occupational therapist.
ties were high, with r = 0.99 and r = 0.91, respectively (n = 13) One of the purposes of this study was to determine new
(Wilson et al., 1987b). The high interrater reliability of the test cutoff scores for the Behavioural Subtest of the CBIT-HK, if
battery and its subtests, and the high parallel-form reliability any, for normality and for patients with stroke and unilateral
of the Behavioural Subtest in this study contribute support for neglect in Hong Kong. The Behavioural Subtest as well as its
the reliable testing procedure and scoring system in the new items was sensitive to the presence of unilateral neglect
Chinese version. The intrarater (test–retest) reliabilities were except for the items of Article Reading, Address and Sentence
a little lower than the original version (r = 0.91, n = 10) (Wilson Copying, and Card Sorting. The findings of lower cutoffs for
et al., 1987b), which might be attributed to the small sample size two language-related items in the CBIT-HK, i.e. Article Reading
(n = 7). However, results also suggested that the Conventional and Address and Sentence Copying, given lower sensitivities
Subtest (r = 0.82) was more reliable than the Behavioural among the items in the Behavioural Subtest, suggests that the
Subtest (r = 0.75) among raters in test–retest administration. use of Chinese words and sentences in these items may not have
Instead of using both subtests and the whole test battery for equivalent sensitivities to the original English version in deter-
assessing patients with unilateral neglect, the Conventional mining neglect phenomenon. Apart from cultural and linguistic
Subtest is suggested to be used individually if the subject can- differences, the Chinese content that was developed on top of
not tolerate a long length of time (30–45 minutes) spent in the the content of the Chinese version of the BIT in Wuhan showed
evaluation. This interpretation is consistent with Wilson et al. lower item–total correlations as compared with the other
(1987b) that the Conventional Subtest might be used to diag- behavioural items in the test (r = 0.56 for Article Reading and
nose the presence or absence of neglect, and the Behavioural r = 0.57 for Address and Sentence Copying). It is likely that
Subtest used to indicate the type of everyday problems that Chinese reading seems to exert a noticeable effect on hamper-
are likely to occur because of neglect, and to guide therapists ing the performance of visual scanning. Previous studies showed
in the selection of tasks to work on in treatment. It is also that visual scanning for Chinese and English is different.
interesting that “Star Cancellation” was the most reliable item Chinese characters are square and rectangle, which by itself
to detect neglect within the whole test battery. Traditionally, does not provide any cue for the reading orientation, whereas
several simple bedside screening tests have been developed for English is presented by alphanumeric transcription of words that
the assessment of neglect, including drawing and copying tasks, are different in length, height and form (Tang, Au Yeung &
line bisection task, and various cancellation tasks. Although Chen, 1997). English language reader has an effective visual
many clinicians are familiar with object copying and clock field or perceptual span that is asymmetric, extending about
drawing tests, these are not very sensitive on their own, and four characters to the left of the letter being fixated and about
are also not always easy to score in a graded manner (Parton, 15 characters to the right (Rayner et al., 1981). Goonetilleke,
Malhotra & Husain, 2004). Furthermore, patients with con- Lau & Shih (2002) found that the cultural difference could
structional dyspraxia may perform poorly on copying tasks, possibly interact with different types of layout and had an
showing copying errors on both the left and the right side of impact on search pattern but not search performance. In the
space, even though they do not show any spatial neglect (Parton study, Hong Kong Chinese had significantly better performance
et al., 2004). For the line bisection task, a paradoxical leftward on the top horizontal area of the search field, while Mainland

Hong Kong Journal of Occupational Therapy 29

30
Kenneth N.K. Fong, et al

Table 5. Spearman’s ␳ correlation between the CBIT-HK and other concurrent measures (n = 54)
Edu MMSE CDT CBIT-C CBIT-B CBIT-T Self-care Sphincter Transfer Locom Com Social Total Motor Total Cog

Edu
MMSE 0.38*
CDT 0.11 0.36†
CBIT-C 0.22 0.46* 0.34†
CBIT-B 0.30† 0.41* 0.27 0.89*
CBIT-T 0.27† 0.47* 0.34† 0.98* 0.95*
Self-care 0.15 −0.03 0.04 0.37* 0.35* 0.38*
Sphincter 0.35* 0.19 −0.01 0.26 0.27 0.27 0.47*
Transfer 0.23 0.03 0.08 0.38* 0.43* 0.41* 0.85* 0.34†
Locom 0.12 0.16 0.10 0.26 0.26 0.26 0.48* 0.28† 0.62*
Com −0.03 0.16 0.10 0.15 0.10 0.15 0.29† 0.24 0.11 0.09
Social 0.16 0.26 0.22 0.46* 0.44* 0.48* 0.40* 0.25 0.30† 0.14 0.50*
Total Motor 0.23 0.00 0.04 0.38* 0.38* 0.40* 0.97* 0.56* 0.90* 0.57* 0.26 0.39*
Total Cog 0.10 0.24 0.18 0.40* 0.35* 0.41* 0.42* 0.31† 0.25 0.16 0.77* 0.93* 0.39*
FIM-T 0.22 0.12 0.09 0.47* 0.44* 0.48* 0.93* 0.55* 0.84* 0.53* 0.45* 0.62* 0.95* 0.64*
*p < 0.01; †p < 0.05. Edu = years of education; MMSE = Mini-Mental State Examination; CDT = Clock Drawing Test; CBIT-C = Conventional Subtest of the CBIT-HK; CBIT-B = Behavioural Subtest
of the CBIT-HK; CBIT-T = total score of the CBIT-HK; Self-care = self-care subscore of the FIM; Sphincter = sphincter subscore of the FIM; transfer = transfer subscore of the FIM; Locom = locomo-
tion subscore of the FIM; Com = communication subscore of the FIM; Social = social subscore of the FIM; Total Motor = total score of the Motor Measure of the FIM; Total Cog = total score of the
Cognitive Measure of the FIM; FIM-T = total score of the FIM.
HKJOT 2007;17(1)

Hong Kong Journal of Occupational Therapy

 RELIABILITY AND VALIDITY OF THE CBIT-HK

Chinese had better search performance on the top left side of In closing, the study had two major limitations. First, Part
the screen. Non-native Chinese readers had no preference on I of the study involved a small sample size, and both samples
any particular area (Goonetilleke et al., 2002). It should be of parts I and II were convenience samples of patients with
remembered that there is no literature thus far on the cultural stroke, recruited from the occupational therapy department of
differences in visual search performance in subjects with stroke a single rehabilitation hospital, which limits the representative-
and unilateral neglect; more work needs to be done on the ness of the sample and, hence, the generalizability of the results.
importance of written representation in Chinese in order to Second, this was a cross-sectional study that did not measure
assess visual scanning difficulty in subjects with unilateral neg- the changes in the severity of unilateral neglect over time, so
lect. On the other hand, the lower sensitivity for Card Sorting that the sensitivity of the CBIT-HK in measuring change in
from our results remains a question. One of the explanations neglect over time was unknown.
could be that the playing cards had to be separated approxi- We established the reliability and validity of the CBIT-HK
mately 2 inches apart, which made the layout too long for the for research purposes, which was designed to assess the uni-
patient to search from top to bottom (i.e. vertical search). lateral neglect or visual inattention of patients with stroke and
Results of concurrent validity indicated that both subtests unilateral neglect. This study will make this tool a unique con-
had similar correlations with the total score of the CBIT-HK tribution to clinical application in occupational therapy as well
as well as the functional subscales as measured by the FIM. as for future study in Hong Kong.
The Behavioural Subtest was a far transfer task explicitly
focusing on real-life activities and therefore should be more Acknowledgements
predictive for real-life situations. However, the lack of differ-
ences between the correlations of the two subtests with self- Thank you to all the participants and staff of the Department
care and transfer suggests that the Behavioural Subtest is not of Rehabilitation and Occupational Therapy, Kowloon Hospital.
superior to the Conventional Subtest in relating to real-life In particular, thanks should be given to Ms Serena Cheng and
functional performance. Moreover, the Conventional Subtest Ms Dora Chan for their administrative support in Kowloon
also had a significant moderate relationship with the CDT, Hospital throughout the study. The authors would also like to
which measures unilateral neglect by drawing a clock face thank Miss Vicky Chun, Ms Ada Chow, and Miss Tang Kwai
in degrees quantitatively, whereas the Behavioural Subtest did Fun for their assistance in data collection.
not. This result was consistent with the findings of Hartman-
Maeir & Katz (1995) in an Israeli sample that some of the References
items in the Behavioural Subtest did not correlate with paral-
lel functional measures. The above finding again supports the Azouvi, P., Oliver, S., de Montely, G., Samuel, C., Louis-Dreyfus, A. &
Tesio, L. (2003). Behavioral assessment of unilateral neglect: study of the
recommendation by the investigators that the Conventional psychometric properties of the Catherine Bergego Scale. Archives of
Subtest can be used individually in neglect measurement, Physical Medicine and Rehabilitation, 84, 51–57.
especially for research purposes. This interpretation was con- Calvanio, R., Levine, D. & Petrone, P. (1993). Elements of cognitive
sistent with Wilson et al. (1987b) that the Conventional rehabilitation after right hemisphere stroke. Behavioral Neurology, 11,
Subtest might be used to diagnose the presence or absence of 25–57.

neglect, and the Behavioural Subtest used to indicate the type Chan, M., Chan, B. & Fong, K. (2004). Reliability and Validity of the
of everyday problems likely to occur because of neglect, and Chinese Behavioral Inattention Test (BIT) (Hong Kong Chinese Version).
Paper presented at the Occupational Therapy Symposium, February
to guide therapists in the selection of tasks to work on in treat- 2004, Hong Kong.
ment. Although the CBIT-HK and both subtests had moderate
Chiu, H.F.K., Lee, H.C., Chung, W.S. & Kwong, P.K. (1994). Reliability
relationship with the MMSE, the findings of significant corre- and validity of the Cantonese version of mini-mental state examination: a
lation of the Behavioural Subtest with education particularly preliminary study. Journal of Hong Kong College of Psychiatry, 4, 25–28.
suggests that the performance of language-related items such Denes, G., Semenza, C. & Stoppa, E. (1982). Unilateral spatial
as Article Reading, and Address and Sentence Copying may neglect and recovery from hemiplegia: a follow-up study. Brain, 105,
543–552.
be influenced by years of education, and the results of the
Behavioural Subtest should be interpreted with caution so as to Feigenson, J.S., McCarthy, M.L. & Greenberg, S.D. (1997). Factors
influencing outcome and length of stay in a stroke rehabilitation unit.
detect neglect manifestations rather than the subject’s literacy. Stroke, 8, 657–662.
The significant relationship with the MMSE suggests that
Goonetilleke, R.S., Lau, W.C. & Shih, H.M. (2002). Visual search strate-
patient’s cognitive performance need to be considered when gies and eye movements when searching Chinese character screens.
interpreting the results of either the CBIT-HK or its subtests. International Journal of Human-Computer Studies, 57, 447–468.

Hong Kong Journal of Occupational Therapy 31

Kenneth N.K. Fong, et al HKJOT 2007;17(1)

Hartman-Maeir, A. & Katz, N. (1994). Validity of the Behavioral reading. Journal of Experimental Psychology: Human Perception and
Inattention Test (BIT): relationships with functional tasks. The American Performance, 7, 169–179.
Journal of Occupational Therapy, 49, 507–516. Rapcsak, S.Z., Verfaelle, M., Fleet, W.S. & Heilman, K.M. (1989).
Heilman, K.M., Watson, R.T. & Valenstein, A. (1985). Neglect and Selective attention in hemispherical neglect. Archives of Neurology, 46,
related disorders. In: Heilman, K.M. & Valenstein, E. (eds.) Clinical neu- 178–182.
ropsychology (2nd ed.), pp. 243–293. New York: Oxford University Press.
Riddoch, M.J. & Humphreys, A.W. (1983). The effect of cueing on uni-
Kaplan, R.F., Verfaelle, M., Meadows, M.E., Caplan, L.R., Pessin, M.S. & lateral neglect. Neuropsychologia, 21, 589–599.
Dewitt, L.D. (1991). Changing attentional demands in left hemispatial
Stone, S.P., Wilson, B., Wroot, A., Halligan, P.W., Lange, L.S., Marshall,
neglect. Archives of Neurology, 48, 1263–1266.
J.C. & Greenwod, R.C. (1991). The assessment of visuo-spatial neglect
Katz, N., Hartman-Maeir, A., Ring, H. & Soroker, N. (1999). Functional after acute stroke. Journal of Neurology, Neurosurgery, and Psychiatry,
disability and rehabilitation outcome in right hemisphere damaged patients 54, 345–350.
with and without unilateral spatial neglect. Archives of Physical Medicine
and Rehabilitation, 80, 379–384. Suhr, J.A. & Grace, J. (1999). Brief cognitive screening of right hemi-
sphere stroke: relation to functional outcome. Archives of Physical
Kinsbourne, M. (1987). Mechanisms of unilateral neglect. In: Jeannerod, M. Medicine and Rehabilitation, 80, 773–776.
(ed.) Advances in psychology 45: neurological and neuropsychological
aspects of spatial neglect, pp. 69–86. Amsterdam: Elsevier Science. Sunderland, A., Derick, T. & Hewer, R.L. (1987). The natural history of
visual neglect after stroke. International Disability Studies, 9, 55–59.
Kinsella, G. & Ford, B. (1980). Acute recovery patterns in stroke patients:
neuropsychological factors. Medical Journal of Australia, 2, 663–666. Tang, C.K., Au Yeung, L.H. & Chen, H.C. (1997). The effective visual
field in Chinese reading. In: Chen, H.C. (ed.) Cognitive processing of
Kotila, M., Niemi, M.L. & Laaksonen, R. (1986). Four year prognosis of
Chinese and related Asian languages, pp. 267–286. Hong Kong: The
stroke patients with visuospatial inattention. Scandinavian Journal of
Chinese University Press.
Rehabilitation Medicine, 18, 177–179.
Ladavas, E., Menghini, G. & Umilta, C. (1994). On the rehabilitation Uniform Data System (UDS) for Medical Rehabilitation (1994). Func-
of hemispatial neglect. In: Riddoch, M.J. & Humphreys, G.W. (eds.) tional Independence Measure training manual. U.S.: U.B. Foundation
Cognitive trunk rotation with neuropsychology and cognitive rehabilitation, Activities.
pp. 151–172. London: Erlbaum. Watson, Y.I., Arfken, C.L. & Birge, S.J. (1993). Clock completion: an
Massironi, M., Antonucci, G., Pizzamiglio, L., Vitale, M.V. & Zoccolotti, P. objective screening test for dementia. Journal of American Geriatric
(1988). The Wundt-Jastrow illusion in the study of spatial hemi-inattention. Society, 41, 1235–1240.
Neuropsychologia, 26, 161–166. Wilson, B., Cockburn, J. & Halligan, P. (1987a). Development of a
Posner, M.J., Walker, J.A., Friedlich, F.J. & Rafal, R.D. (1987). How do behavioral test of visuospatial neglect. Archives of Physical Medicine and
the parietal lobes direct covert attention. Neuropsychologia, 25, 135–145. Rehabilitation, 68, 98–102.
Rayner, K., Inhoff, A.W., Morrison, R.E., Slowwiaczer, M.L. & Bertera, J.H. Wilson, B., Cockburn, J. & Halligan, P. (1987a). Behavioural Inattention
(1981). Masking of foveal and parafoveal vision during eye fixations in Test manual. Suffolk: Thames Valley Test Company.

32 Hong Kong Journal of Occupational Therapy

 RELIABILITY AND VALIDITY OF THE CBIT-HK

Appendix. Record form of the CBIT-HK

`ÜáåÉëÉ=_ÉÜ~îáçìê~ä=få~ííÉåíáçå=qÉëí
(Please affix label)

Hospital / OPD No:
eçåÖ=hçåÖ=îÉêëáçå=E`_fqJehF
ElÅÅìé~íáçå~ä=qÜÉê~éóI=hçïäççå=eçëéáí~äF
Name:

 !"#= E
 !"F
E
 !"#$%&'()*F ID No: Sex: Age:

^
 !"#$%&'()

 !
 !

 !
 !"

 !
 !"

 !"
 !"= = = = = = = = = = = = = = =


 !
 !"#= = = = = = = = = = = = = = =


 !"
 !"#$= = = = = = = = = = = = =


 W= ^= = = = = = = = = = = = = = = = = _
=
=


 !
 !

 !


 !


 !












 !"#$%

 !"#

_
 !"#$

 !"#

 = =

 !"#

 = = =

N
 LPS PQ
L
 N
 !" LV R
L

O
 ! LQM PO
L
 O
 ! LV T
L

P
 LRQ RN
L
 P
 ! LV U
L

Q
 !"#$% LQ P
L
 Q
 !" LV R
L

R
 !" LV T
L
 R
 !"# $ LV U
L

S
 !"# LP O
L
 S
 !" LV U
L

= = = = = = = = =
 LNQS NOV
L
 T
 !"#$% LV Q
L

U
 ! LV U
L

V
 !" LV U
L


 LUN SN
L


 !"W

Hong Kong Journal of Occupational Therapy 33

CBIT 33 7/2/07, 4:44 PM

View publication stats